$Intel(INTC)$ The global AI accelerator market is over $200B a year, with roughly two-thirds of that going to inference. Nvidia and AMD are both constrained by HBM3e and TSMC CoWoS supply. Intel's D270 takes a different route with a compact Xe3P die and LPDDR5X memory up to 480GB, which removes the HBM bottleneck entirely and should support 60%+ gross margins. If Intel manages to capture just 5% to 10% of the air-cooled enterprise inference market in 2027-2028, that works out to $10B to $20B+ a year in high-margin GPU revenue, plus billions more in attached Xeon host CPUs for every 8-GPU server.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
Comments